[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-94904":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":10,"language":11,"languages":10,"totalLinesOfCode":10,"stars":12,"forks":13,"watchers":13,"openIssues":14,"contributorsCount":14,"subscribersCount":14,"size":14,"stars1d":14,"stars7d":15,"stars30d":15,"stars90d":14,"forks30d":14,"starsTrendScore":16,"compositeScore":17,"rankGlobal":10,"rankLanguage":10,"license":18,"archived":19,"fork":19,"defaultBranch":20,"hasWiki":19,"hasPages":19,"topics":21,"createdAt":10,"pushedAt":10,"updatedAt":38,"readmeContent":39,"aiSummary":40,"trendingCount":14,"starSnapshotCount":14,"syncStatus":41,"lastSyncTime":42,"discoverSource":43},94904,"procedural-sounds","m1ckc3s\u002Fprocedural-sounds","m1ckc3s","Procedural interface sounds, with taste. Generated live in the browser from recipes, steered by a hand-curated library and calibrated by ear. No audio files. Beta, open to collaborators.","https:\u002F\u002Fprocedural-sounds.vercel.app",null,"TypeScript",149,3,0,27,1,58.01,"MIT License",false,"main",[22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37],"design-tools","generative-audio","interface-design","microinteractions","nextjs","open-source","procedural-audio","react","sound-design","synthesizer","tailwindcss","typescript","ui-sounds","ux","web-audio","web-audio-api","2026-08-24 04:01:22","# Procedural interface sounds, with taste\n\nThis project does not have a settled name yet. Naming ideas are welcome, open an issue.\n\nEvery sound is synthesized live in the browser from a recipe. There are no audio files\nanywhere in the product. Pick a category, generate until one fits, hear it in context,\nexport it as a WAV or as a few lines of JavaScript.\n\n## Where this is at\n\n**Early beta, and very much a work in progress.** The core loop works: generate, hear it in\ncontext, export. The library is growing, the generators are learning, and a lot of what is\nhere is the first version of itself. Do not read this repo as a finished project. Read it as\none you can help build.\n\nI am looking for help, and PRs are welcome. In rough order of how much they would matter:\n\n- **Installation and packaging.** Today the only way to use a sound is to export it from the\n  running app. An npm package that ships the player and lets a project `import` a sound is\n  the single biggest missing piece. The player is already standalone and dependency-free.\n- **Training data.** The library grows by curation, and every keep or delete trains the\n  generators. If you have ears and patience this is the most valuable thing anyone can do.\n- **The generators.** New instruments, figures, archetypes and characters, and better\n  training for the ones that exist.\n- **The front end.** The stage animations, the product page, the workbench.\n- **The docs.** Anything in `docs\u002F` that is stale or unclear, and the Atlas copy.\n\n[CONTRIBUTING.md](CONTRIBUTING.md) has the specifics, including how curation contributions\nget merged. [`docs\u002FTODO.md`](docs\u002FTODO.md) is the live work list.\n\n## Run it\n\nNode 24 (Next 16, React 19, Tailwind 4).\n\n```bash\nnpm install\nnpm run dev\n```\n\nOpen the product at `\u002F`. At the bottom of the page there is a **dev mode** link. That is\nthe workbench, and it only exists when you run locally: its routes write curation state to\n`data\u002Fpool\u002F*.json` on your filesystem, and a production build closes them.\n\nInside the workbench you can:\n\n- **Library**: browse everything, sort new keeps into categories by ear, and audit the\n  library against the ear-safety ceilings.\n- **Variations, Creations, Invent, Wild, Craft, Prospect**: generate in batches from each\n  engine, keep what is good, delete what is not. Every verdict on Creations and Invent moves\n  that engine's dice. Keeps from every tab land in the to-sort inbox with no category until\n  you sort them.\n- **Editor and Import**: hand-write a recipe, or paste one back in.\n- **Calibrate**: set the ear-safety ceilings and loudness targets by listening.\n- **Atlas**: everything every generator can say, rendered live from the code and the\n  current training state.\n- **Dedupe**: whole-library similar-pair triage. Identical-only; when unsure, keep both.\n- **Trash**: everything deleted or marked duplicate. Restore lives only here.\n\nRead [`docs\u002FTRAINING.md`](docs\u002FTRAINING.md) before a curation session. It says what\neach tab's keep and delete actually write.\n\n## How it works\n\nThe short version: a hand-curated library, a handful of generators that draw from and remix\nit, and a learning layer that shifts each generator's odds every time a curator keeps or\ndeletes what it made. The library and the training data are the same asset.\n\nOn provenance: the synth core began as an adaptation of an open source recipe player, and a\nseveral open source sound packs seeded the very first library. Everything that generates,\nlearns and curates is original to this project, as are the instruments, figures and\ncharacters the generators draw from, and the library has been growing past its seeds ever\nsince. The specific projects and their licenses are in\n[THIRD-PARTY-NOTICES.md](THIRD-PARTY-NOTICES.md), also rendered at `\u002Flicenses`.\n\nThe long version lives in [`docs\u002F`](docs\u002F):\n\n- [GETTING-STARTED.md](docs\u002FGETTING-STARTED.md) sets up the workbench and explains each tab.\n- [HOW-IT-WORKS.md](docs\u002FHOW-IT-WORKS.md) is the explainer, with a glossary.\n- [ARCHITECTURE.md](docs\u002FARCHITECTURE.md) is what is built right now.\n- [CLAUDE.md](.claude\u002FCLAUDE.md) is the rulebook the whole project holds to.\n\n## License\n\nMIT. See [LICENSE](LICENSE).\n","这是一个在浏览器中实时生成界面音效的开源工具，不依赖预录音频文件，所有声音均通过算法合成。核心功能包括基于配方（recipe）的程序化音频生成、人工听觉校准的声库管理、上下文试听、WAV\u002FJS代码导出，以及支持协作式声库训练与分类。技术上基于Web Audio API，使用TypeScript构建，集成React与Next.js，强调设计导向与听觉敏感性。适用于UI\u002FUX设计师、前端开发者及声音设计师快速创建符合品牌调性的微交互音效，尤其适合需要可定制、轻量级、无版权音频资源的Web应用原型与生产环境。",2,"2026-08-18 02:30:03","CREATED_QUERY"]